Our present knowledge of the genes involved insusceptibilities to complex human diseases such as cancer is limited due to the large number of genetic and environmental factors involved and their complex interactions. Cancer is characterized by aberrant signaling through pathways required for regulated cell growth, death, or DNA maintenance. Single gene mutations in these pathways cause rare cancers, but common polymorphic alleles of unknown genes with modest effects on signaling are responsible for most cancers. To find these genes, human studies are difficult, costly, and biased toward previously known biological mechanisms and genes of large effect. An alternative approach is to find modifiers of conserved signaling pathways in model organismsand then study their human homologs variants associated with disease. The Hsp90 stress protein supports the activities of many conserved tumor suppressors, oncogenes, and cell cycle regulators. When Drosophila Hsp90 is limiting, polygenic variation with potent effects on development is expressed as strain specific abnormalities. We believe that the developmental novelties result from natural genetic variation affecting the strength of signaling through the developmental, regulatory, and growth control pathways determined by Hsp90 targets. Thus, a network of natural variation influencingHsp90-dependent signaling pathways in flies models the architecture of signaling variation for cancer predisposition in humans, and has the potential to drive the evolution of development. To identify natural Hsp90-buffered polymorphisms and place them in pathways we used laboratory selection for Hsp90-dependent eye abnormalities to generate replicate 'deformed eye'lines with a high fraction of affected flies. An inbred wild-type genetic background with normal eyes was crossed into the selection lines to construct over 1,400 recombinant isogenic mapping lines (clones). For each recombinant line, on average 50 flies having identical recombinant genotypes were scored for trait penetrance (probability affected) and saved for genotyping. Our specific aims are to 1) localize deformed eye polymorphisms by SQTL (Suppressor of Quantitative Trait Loci) and deletion mapping to test the idea that varied genetic architectures for deformed eye arose in the replicate lines. 2) Clone the wild-type alleles of deformed eye polymorphisms as suppressors of trait penetrance to determine whether hidden polymorphisms reside in Hsp90 targets or in peripheral genes of small effect. 3) Identify developmental processes and pathways responsible for the pathology of deformed eye to bring Hsp90-buffered variation into biological context. We expect this research will identify conserved Hsp90 target pathways and an extended network of naturally polymorphic genes that influence their function.